Cut region of interest images from 3D stack with known indices and without for loop

Hi! I am having trouble cutting region of interest from a 3D stack.

A small example, I have a 3x3x2 array of:

array = [1,2,3;4,5,6;7,8,9];
array(:,:,2) = [1,2,3;4,5,6;7,8,9];

mini image indices:

left = [1,2]
right = [2,3]
top = [1,2]
bottom = [2,3]

create mini images:

miniImages = array(top:bottom,left:right,:)

result:

miniImages(:,:,1) = [1,2;4,5]
miniImages(:,:,2) = [1,2;4,5]

However, I want the result to be:

miniImages(:,:,1) = [1,2;4,5]
miniImages(:,:,2) = [5,6;8,9]

It seems that it only uses the first indices in variables left, right, top, bottom to create all mini images.

Anyone know how to solve this? Currently I use a for loop over the indices, but I want to make it faster.

Thanks!

[Matt J]

miniImages(:,:,2) = [5,6;7,8]

Don't you mean [5 6;8;9] ?

[Kevin Jansen]

Yes! That was a mistake, I will try to edit it.

Answer 1

[Matt J]

I doubt anything will be faster than a loop, but you can try this:

array = [1,2,3;4,5,6;7,8,9];
array(:,:,2) = [1,2,3;4,5,6;7,8,9]

array = 
array(:,:,1) =

     1     2     3
     4     5     6
     7     8     9


array(:,:,2) =

     1     2     3
     4     5     6
     7     8     9

left = [1,2];
top = [1,2];
bottom = [2,3];
right = [2,3];

args=cellfun(@(z)reshape(z,1,1,[])    , {left,top,bottom,right} ,'un' ,0);
[L,T,B,R]=deal(args{:});

[m,n,p]=size(array);

[I,J]=ndgrid(1:m,1:n);

lidx=T<=I & I<=B & L<=J & J<=R;

miniImages=reshape( array(lidx)  ,B(1)-T(1)+1,R(1)-L(1)+1,[])

miniImages = 
miniImages(:,:,1) =

     1     2
     4     5


miniImages(:,:,2) =

     5     6
     8     9

[Kevin Jansen]

Hey, awesome it worked! Much more difficult than I thought it would be. My code with larger arrays ran 380x faster without the for loop. Thank you so much!